基於Model-Agnostic Meta-Learning方法的深度神經網路識別多變量製程中的異常來源

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：67

、訪客IP：3.15.139.250

姓名

林明龍(Ming-Long Lin) 查詢紙本館藏

畢業系所

工業管理研究所

論文名稱

基於Model-Agnostic Meta-Learning方法的深度神經網路識別多變量製程中的異常來源
(Identifying the Sources of Out-of-control Signals in Multivariate Process Using Deep Neural Networks Based on Model-Agnostic Meta-Learning Approach)

相關論文

★ 應用失效模式效應分析於產品研發時程之改善	★ 服務品質因子與客戶滿意度關係研究-以汽車保修廠服務為例
★ 家庭購車決策與行銷策略之研究	★ 計程車車隊派遣作業之研究
★ 電業服務品質與服務失誤之探討-以台電桃園區營業處為例	★ 應用資料探勘探討筆記型電腦異常零件-以A公司為例
★ 車用配件開發及車主購買意願探討(以C公司汽車配件業務為實例)	★ 應用田口式實驗法於先進高強度鋼板阻抗熔接條件最佳化研究
★ 以層級分析法探討評選第三方物流服務要素之研究-以日系在台廠商為例	★ 變動良率下的最佳化批量研究
★ 供應商庫存管理架構下運用層級分析法探討供應商評選之研究-以某電子代工廠為例	★ 台灣地區快速流通消費產品銷售預測模型分析研究–以聯華食品可樂果為例
★ 競爭優勢與顧客滿意度分析以中華汽車為例	★ 綠色採購導入對電子代工廠的影響-以A公司為例
★ 以德菲法及層級分析法探討軌道運輸業之供應商評選研究–以T公司為例	★ 應用模擬系統改善存貨管理制度與服務水準之研究-以電線電纜製造業為例

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

[檢視]

[下載]

本電子論文使用權限為同意立即開放。
已達開放權限電子全文僅授權使用者為學術研究之目的，進行個人非營利性質之檢索、閱讀、列印。
請遵守中華民國著作權法之相關規定，切勿任意重製、散佈、改作、轉貼、播送，以免觸法。

摘要(中)

在現代製造業中，多變量管制圖已被廣泛應用於檢測製程異常。然而，它卻無法直接辨識異常來源，這是製程監控中的一大挑戰。因此，本研究旨在發展一套有效且即時的方法，以應對製造過程中的維度災難，同時識別多變量製程中的異常來源，提高監控的準確性和效率。
資料前處理採用B4-Broken-stick變數選擇方法，有效降低多維度數據的複雜性。本研究使用基於Model-Agnostic Meta-Learning方法的深度神經網路進行異常來源的分類。實驗結果顯示，本研究提出的 MAML of DNNs 模型透過少量樣本學習，快速適應新任務，在不同偏移情況下能夠有效處理所有異常狀況並保持良好的準確率。與 MLP 模型相比，MAML of DNNs 模型在準確率和訓練時間上均表現較佳，證實其在識別多變量製程異常來源方面更具效率和效果。將MAML of DNNs模型應用於生產過程中，可以即時辨識異常來源，幫助相關領域人員更有效率和高效地進行品質改善工作。

摘要(英)

In modern manufacturing, multivariate control charts have been widely used to detect process anomalies. However, they cannot directly identify the source of out-of-control signals, posing a significant challenge in process monitoring. This study aims to develop an effective and real-time method to address the curse of dimensionality in manufacturing processes while identifying the source of out-of-control signals in multivariate processes, thereby improving monitoring accuracy and efficiency.
This study employs the B4-Broken-stick variable selection method to effectively reduce the complexity of high-dimensional data. We propose using a deep neural network based on Model-Agnostic Meta-Learning (MAML) to classify the sources of out-of-control signals. Experimental results show that the proposed MAML of DNNs model learns from a small number of samples and quickly adapts to new tasks, effectively handling all anomaly conditions and maintaining high accuracy under various shift scenarios. Compared to the MLP model, the MAML of DNNs model performs better in both accuracy and training time, demonstrating its efficiency and effectiveness in identifying the sources of out-of-control signals in multivariate processes. Applying the MAML of DNNs model in the production process enables real-time identification of the sources of out-of-control signals, assisting personnel in related fields to conduct quality improvement work more efficiently and effectively.

關鍵字(中)

★ 多變量製程
★ 多變量管制圖
★ 變數選擇
★ 深度神經網路
★ Model-Agnostic Meta-Learning

關鍵字(英)

★ Multivariate Processes
★ Multivariate Control Charts
★ Variable Selection
★ Deep Neural Networks
★ Model-Agnostic Meta-Learning

論文目次

摘要 i
Abstract ii
目錄 iii
圖目錄 v
表目錄 vi
第一章、緒論 1
1.1研究背景與動機 1
1.2研究目的 1
1.3研究範圍與假設 2
1.4論文架構 2
第二章、文獻探討 4
2.1統計製程管制 4
2.1.1多變量統計製程管制 5
2.1.2 Hotelling T2管制圖 7
2.2識別異常來源相關文獻 9
2.2.1傳統統計方法 9
2.2.2機器學習方法 10
2.3主成分分析 (Principal Component Analysis, PCA) 11
2.3.1 B4-Broken-stick方法 11
2.4神經網路與深度學習 12
2.4.1神經網路 (Neural Networks, NNs) 13
2.4.2深度學習 (Deep Learning, DL) 14
2.4.3神經網路在統計製程管制上之應用 15
2.5元學習 (Meta-Learning) 16
2.5.1 Model-Agnostic Meta-Learning (MAML) 18
第三章、研究方法 19
3.1研究架構 19
3.2原始資料 20
3.3使用B4-Broken-stick方法進行變數選擇 21
3.4生成隨機多變量過程 23
3.4.1異常資料 23
3.5模型架構 24
3.5.1輸入特徵向量 24
3.5.2隱藏層 25
3.5.3輸出層 25
3.6 Model-Agnostic Meta-Learning of Deep Neural Networks 25
3.6.1問題及符號定義 26
3.6.2 MAML演算法 26
3.6.3應用於多標籤二元分類 28
第四章、實驗結果與分析 30
4.1資料前處理 30
4.2實驗設計 30
4.2.1元訓練集和元測試集之資料與樣本數 30
4.2.2最佳化模型配置 31
4.3實驗分析與評估 32
4.3.1分類結果 32
4.3.2模型比較 34
4.3.3文獻案例之模型比較 36
第五章、結論與未來研究方向 40
參考文獻 41

參考文獻

[1] Abdi, H., & L. J. Williams. "Principal component analysis." Wiley Interdisciplinary Reviews: Computational Statistics 2.4, 2010, 433-459.
[2] Aslam, M., A. Saghir, & L. Ahmad. Introduction to Statistical Process Control. John Wiley & Sons, 2020.
[3] Bersimis, S., S. Psarakis, & J. Panaretos. "Multivariate statistical process control charts: an overview." Quality and Reliability Engineering International 23.5, 2007, 517-543.
[4] Chen, L. H., & T. Y. Wang. "Artificial neural networks to classify mean shifts from multivariate χ2 chart signals." Computers & Industrial Engineering 47.2-3, 2004, 195-205.
[5] Cheng, C. S. "A multi-layer neural network model for detecting changes in the process mean." Computers & Industrial Engineering 28.1, 1995, 51-61.
[6] Cheng, C. S., & H. T. Lee. "Diagnosing the variance shifts signal in multivariate process control using ensemble classifiers." Journal of the Chinese Institute of Engineers 39.1, 2016, 64-73.
[7] Diren, D. D., S. Boran, & I. Cil. "Integration of machine learning techniques and control charts in multivariate processes.", 2020.
[8] Doganaksoy, N., F. W. Faltin, & W. T. Tucker. "Identification of out of control quality characteristics in a multivariate manufacturing environment."
Communications in Statistics-Theory and Methods 20.9, 1991, 2775-2790.
[9] Finn, C., P. Abbeel, & S. Levine. "Model-agnostic meta-learning for fast adaptation of deep networks." International Conference on Machine Learning. PMLR, 2017, 1126-1135.

[10] Frontier, S. "Étude de la décroissance des valeurs propres dans une analyse en composantes principales: Comparaison avec le modd́le du bâton brisé." Journal of Experimental Marine Biology and Ecology 25.1, 1976, 67-75.
[11] Guh, R. S., & Y. R. Shiue. "An effective application of decision tree learning for on-line detection of mean shifts in multivariate control charts." Computers & Industrial Engineering 55.2, 2008, 475-493.
[12] Guo, Y., & K. J. Dooley. "Identification of change structure in statistical process control." The International Journal of Production Research 30.7, 1992, 1655-1669.
[13] Hawkins, D. M. "Multivariate quality control based on regression-adiusted variables." Technometrics 33.1, 1991, 61-75.
[14] Hawkins, D. M. "Regression adjustment for variables in multivariate quality control." Journal of Quality Technology 25.3, 1993, 170-182.
[15] Hayter, A. J., & K. L. Tsui. "Identification and quantification in multivariate quality control problems." Journal of Quality Technology 26.3, 1994, 197-208.
[16] Jackson, D. A. "Stopping rules in principal components analysis: a comparison of heuristical and statistical approaches." Ecology 74.8, 1993, 2204-2214.
[17] Jackson, J. E. "Multivariate quality control." Communications in Statistics-Theory and Methods 14.11, 1985, 2657-2688.
[18] Jolliffe, I. T. "Discarding variables in a principal component analysis. I: Artificial data." Journal of the Royal Statistical Society Series C: Applied Statistics 21.2, 1972, 160-173.
[19] King, J. R., & D. A. Jackson. "Variable selection in large environmental data sets using principal components analysis." Environmetrics: The Official Journal of the International Environmetrics Society 10.1, 1999, 67-77.

[20] Krenker, A., J. Bešter, & A. Kos. "Introduction to the artificial neural networks."
Artificial Neural Networks: Methodological Advances and Biomedical Applications. InTech, 2011, 1-18.
[21] Mason, R. L., N. D. Tracy, & J. C. Young. "Decomposition of T 2 for multivariate control chart interpretation." Journal of Quality Technology 27.2, 1995, 99-108.
[22] Montgomery, D. C. Introduction to Statistical Quality Control. John Wiley & Sons, 2019.
[23] Murphy, B. J. "Selecting out of control variables with the T2 multivariate quality control procedure." Journal of the Royal Statistical Society Series D: The Statistician 36.5, 1987, 571-581.
[24] Navamani, T. M. "Efficient deep learning approaches for health informatics."
Deep Learning and Parallel Computing Environment for Bioengineering Systems. Academic Press, 2019, 123-137.
[25] Niaki, S. T. A., & B. Abbasi. "Fault diagnosis in multivariate control charts using artificial neural networks." Quality and Reliability Engineering International
21.8, 2005, 825-840.
[26] Noorossana, R., M. Farrokhi, & A. Saghaei. "Using neural networks to detect and classify out‐of‐control signals in autocorrelated processes." Quality and Reliability Engineering International 19.6, 2003, 493-504.
[27] Psarakis, S. "The use of neural networks in statistical process control charts."
Quality and Reliability Engineering International 27.5, 2011, 641-650.
[28] Song, H., Q. Xu, H. Yang, & J. Fang. "Interpreting out-of-control signals using instance-based Bayesian classifier in multivariate statistical process control."
Communications in Statistics-Simulation and Computation 46.1, 2017, 53-77.
[29] Vanschoren, J. "Meta-learning: A survey." arXiv preprint arXiv:1810.03548, 2018.

[30] Western Electric Company. Statistical Quality Control Handbook. AT & T Technologies, 1958.
[31] Wold, S., K. Esbensen, & P. Geladi. "Principal component analysis."
Chemometrics and Intelligent Laboratory Systems 2.1-3, 1987, 37-52.
[32] Yu, J., L. Xi, & X. Zhou. "Identifying source (s) of out-of-control signals in multivariate manufacturing processes using selective neural network ensemble."
Engineering Applications of Artificial Intelligence 22.1, 2009, 141-152.
[33] Zorriassatine, F., & J. D. T. Tannock. "A review of neural networks for statistical process control." Journal of Intelligent Manufacturing 9, 1998, 209-224.

指導教授

葉英傑(Ying-Chieh Yeh)

審核日期

2024-7-17

推文